Introduction
Depression is a global health problem (World Health Organization
2008), and thus new treatments need to translate worldwide. A major barrier to treating depression is limited access to psychological therapies, and the development of novel psychological interventions is a crucial area for research (Wittchen
2012). Cognitive science offers a means for the development of such new interventions via the identification of key processes involved in the maintenance of disorders and the means to modify these. A body of research has demonstrated that depression is characterized by negative biases in many aspects of processing including memory, attention, and interpretation (e.g. Mathews and MacLeod
2005; Gotlib and Joormann
2010; Everaert et al.
2012). Paradigms that aim to directly modify such biases, referred to as “cognitive bias modification” (CBM; MacLeod et al.
2009; Hertel and Mathews
2011) hold promise as novel accessible interventions for depression. However, the investigation of CBM for depression is in its infancy, with research tending to focus on anxiety (see e.g. MacLeod and Mathews
2012). CBM research in depression has also been largely restricted to European and English-speaking countries. If CBM paradigms are to help tackle a global problem such as depression then they need have global applicability. Thus a key question at this early stage of clinical research is whether a CBM paradigm developed in one setting can be successfully translated to and applied in a different country and culture.
A CBM paradigm that has shown some early promise as a potential intervention in depression is an adapted version of the interpretation training paradigm originally developed by Mathews and Mackintosh (
2000). Participants are repeatedly presented with scenarios that start ambiguous, but are then resolved positively, with the aim of training a bias to automatically expect positive resolutions for novel ambiguous situations. Depression is characterised by the tendency to interpret ambiguous information negatively, a negative interpretation bias (Butler and Mathews
1983), and thus such training may be beneficial. In adapting the interpretation CBM (CBM-I) to depression there has been a particular focus on the use of mental imagery (Holmes et al.
2009a). Participants are required to imagine themselves in the scenarios presented, “as if actively involved, seeing them through your own eyes”, and early experimental studies have demonstrated the crucial role of this use of imagery in the effects of the CBM-I paradigm over a single session of training in non-clinical samples (Holmes et al.
2006; Holmes et al.
2008a,
2009b). However, when applied to depression the requirement to
imagine the positive resolutions of the training scenarios may have even greater importance in reducing the clinical symptoms of the disorder.
Depression is characterised by a deficit in positive future imagery (Holmes et al.
2008b; Morina et al.
2011), such that people with depression may struggle to imagine anything other than negative possibilities in their future. Repeated practice in generating positive mental imagery may therefore be particularly helpful. In addition, depression is characterised by a bias for a verbal, ruminative style of processing (Koster et al.
2011), and a bias for observer (seeing oneself from the outside) perspective imagery (Williams and Moulds
2007; Nelis et al.
2013), and thus the particular emphasis in the training on using field perspective imagery and avoiding verbal analysis may be especially useful for positive outcomes in depression. However, no study to date has explored the importance of the instruction to imagine the training scenarios on clinical outcomes in individuals with depression, and none have examined whether practice in imagery use increases imagery vividness in this population.
Two studies have so far investigated the potential of imagery CBM-I in reducing symptoms of depression in clinical samples, when delivered as a stand-alone intervention. In the first such translational study, Blackwell and Holmes (
2010) used a single case series design to investigate the impact of 1 week of daily sessions of imagery CBM-I in seven participants experiencing a current major depressive episode. This study demonstrated the initial promise of imagery CBM-I as a potential intervention in depression, with the group overall showing large effect sizes for improvements in depressive symptoms, interpretive bias, and general mental health at 1-week post-intervention. Depressive symptoms were also measured at a 2-week follow-up and the improvements were maintained. A main limitation of this initial study was the lack of a control group, meaning that improvement may have been attributed to non-specific aspects of the CBM-I such as distraction.
A second study therefore compared the imagery CBM-I to a control condition (Lang et al.
2012). This study used a “multi-component” CBM, comprising three sessions of the paradigm described above, two sessions of a picture-word imagery CBM-I paradigm (Holmes et al.
2008c; Pictet et al.
2011), and one session of a CBM targeting appraisals of negative intrusive memories (Lang et al.
2009). Twenty-six participants with major depression were randomly assigned to complete either 1 week of daily sessions of the positive imagery CBM-I, or a control program. In the control program half of the training stimuli were resolved positively and half were resolved negatively, thus removing the training contingency to always expect a positive resolution. Individuals receiving the positive imagery CBM-I demonstrated significant improvements from pre to post-treatment in depressive symptoms, cognitive bias, and intrusive symptoms, compared to the control condition. Improvements in depressive symptoms at 2-week follow-up were at trend level compared to the control condition. This study therefore provided further support for the clinical potential of imagery CBM-I in depression. It also demonstrated the importance of the consistently positive resolution of the training materials, rather than the effects of the program being due simply to generation of imagery per se, or non-specific effects such as distraction.
These two initial studies pave the way for larger clinical trials investigating the effects of imagery CBM-I over a longer time period. However, they also leave unanswered two key questions about the imagery component of the interventions. First, is the requirement to generate mental imagery crucial for the clinical impact of the training? In the study by Lang et al. (
2012) both conditions involved generating imagery. The superiority of the positive condition suggests that it is practising
positive imagery, rather than imagery per se, that leads to clinical improvement, despite recent research showing a general reduction in imagery vividness in depression (Torkan et al.
2012). However, this study cannot rule out the possibility that in a clinical sample of depressed individuals, simple repeated exposure to hundreds of positive-valenced stimuli may act as a positive mood induction (e.g. Velten
1968) and thus lead to improvements in symptoms of depression. We would in fact predict that without the requirement to generate imagery, people with depression would revert to a verbal, comparative, style of processing and thus fail to show improvement (cf. Holmes et al.
2009b). However, given the importance placed on generating imagery in the paradigm, the role of the instruction to use imagery is a crucial one to be tested.
Another question is whether the repeated practice in generating mental imagery leads to changes in imagery ability. This is a key question for understanding how engaging in imagery CBM-I leads to clinical benefits, as we would hypothesise that part of the helpful effects may be conferred by improvement in the ability to vividly imagine positive imagery. However, it is unclear whether imagery vividness can in fact be improved via training (Rademaker and Pearson
2012), and thus demonstrating a simple training effect on imagery vividness is a crucial first step in testing our hypothesis.
The current study therefore had three general aims. The primary aim was to investigate whether the imagery CBM-I could be successfully translated to and applied in a new, non-European and non-English-speaking, population and culture (Iran). A second aim was to investigate whether the requirement to imagine the positive training materials was important for the clinical impact of the training. A third aim was to investigate whether the training had any impact on the general ability to generate vivid mental imagery.
We addressed the primary aim by investigating the imagery CBM-I in a sample of treatment-seeking patients with major depression in outpatient psychiatry clinics in Iran. The initial theoretical and experimental work underpinning this imagery CBM-I approach was largely carried out in European settings, and although there are no specific theoretical reasons to expect that the approach would be ineffective in a non-European population or culture, a straightforward equivalence of effects cannot be assumed. With some exceptions (e.g. Memory Specificity Training for depression; Neshat-Doost et al.
2013) CBM interventions that showed initial promise, e.g. attention bias modification for social anxiety disorder, have sometimes struggled to translate from one language, country or platform to another (e.g. Carlbring et al.
2012). Research in mental health may be particularly sensitive to translational issues, due to cross-cultural differences in the conception and expression of psychological disorders, as well as in variation in language. In Iran there appears to be a particularly high expression of somatic symptoms in major depression (Hakimshooshtary et al.
2007), which may suggest differences in underlying processes and potential responses to treatments. For example, it has been suggested that Behavioural Activation may be a particularly effective treatment for depression in Iran due to the approach fitting well with Iranian culture (Moradveisi et al.
2013). Thus, demonstrating that an approach can survive translation from its initial place of development is an important step in developing interventions of broader potential reach. The sample in the current study represents a novel group not only in the translation to a different language and culture, but also a “real-world” sample of treatment-seeking patients. With some exceptions (e.g. Brosan et al.
2011), very few studies have investigated CBM in individuals in psychiatric settings. As a first translation to this novel population, we conducted this as a pilot study. In order to enhance comparability with published studies, we used a time-frame for the study as in the previous initial studies, namely a 1-week intervention and 2-week follow-up (Blackwell and Holmes
2010; Lang et al.
2012). We investigated the impact of the imagery CBM-I on symptoms of depression, anxiety, and negative interpretive bias (cf. Lang et al.
2012).
In order to investigate the importance of the imagery instructions, we included a control condition (“non-imagery”) in which participants were presented with an identical set of training stimuli (i.e. all positively resolved) over an identical schedule of sessions, but were given no training in imagery or instruction to imagine the scenarios. Instead they were told to listen to the scenarios and not instructed to use any particular form of processing. We further included an additional no treatment control condition in order to provide a comparison for the active control condition (cf. Watkins et al.
2009).
Finally, in order to investigate the effects of the training on mental imagery, we included at pre-treatment, post-treatment and follow-up a general measure of vividness of mental imagery in order to investigate whether the repeated practice in using mental imagery led to increased imagery vividness. We also included measures of general use of imagery and tendency to ruminate (a form of verbal processing) in everyday life in order to more fully characterise any effects of the CBM-I on imagery and verbal processing. For example, it may be that the repeated requirement to use imagery and not verbal processing within the task would generalize to a greater tendency to use mental imagery and reduced tendency to use a verbal (ruminative) processing style in everyday life. Although these are not hypothesised to be key mechanisms by which the CBM-I paradigm improves clinical outcomes, understanding the broader potential impact of the program is helpful in describing its effects more fully.
We hypothesised that:
1.
The imagery CBM-I would be successfully translated to and applied in the new population, such that participants in the imagery condition would show improvements in symptoms of depression and anxiety, and reduction in negative interpretive bias over the 1-week training and at 2-week follow-up.
2.
Participants in the imagery condition would show greater improvement in outcome measures than those in the non-imagery active control condition.
3.
Participants in the imagery condition would show increased vividness of mental imagery following the imagery CBM-I compared to those in the control conditions.
Discussion
The present pilot study is, to our knowledge, the first investigation of imagery CBM-I for depression in a non-European, non-English-speaking country, the first to investigate the importance of the use of imagery instructions in the clinical impact of repeated sessions of this paradigm, and the first to investigate whether the training can increase the vividness of imagination in depressed individuals. In a sample of patients presenting with major depression to psychiatric outpatient clinics in Iran, compared to two control conditions we found that engaging in repeated sessions of imagery CBM-I over the course of 1 week reduced symptoms of depression and negative interpretive bias at 1-week post-treatment and 2-week follow-up. A significant reduction in trait anxiety was also found at 2-week follow-up. Further, we found evidence supportive of the importance of the instruction to imagine the training scenarios in the clinical impact of the training. In fact, participants simply instructed to listen to and focus on the scenarios showed no more improvement in symptoms of depression or other outcomes than a no treatment control group. Finally, we found that repeated practice in generating imagery resulted in increases in general (non-emotional) mental imagery ability, as measured by self-reported imagery vividness. This study therefore provides preliminary evidence for the potential cross-cultural applicability of a positive imagery training paradigm in depression, and furthers our knowledge of the parameters and effects of this form of training. The implications of the current study and how it builds on previous work will be discussed first in relation to the clinical outcomes, then the role of imagery in the training, and finally the impact of the training on imagery.
The success of the imagery CBM-I in the current study when translated to a new setting and population is encouraging, and supplements the initial findings from other preliminary studies to date (Blackwell and Holmes
2010; Lang et al.
2012). Consistent with these studies, which also used a schedule of 1 week of daily imagery CBM-I and a 2-week follow-up, significant reductions in symptoms of depression and negative cognitive bias were found over the 1-week intervention (
n = 13 per group) and subsequent 2 week follow-up (
n = 8 per group), corresponding to large effect sizes. Rates of clinically significant change in symptoms of depression were generally high in this study in comparison to previous studies. For example, in the imagery condition, the percentage of participants showing clinically significant change were 69.2 and 87.5 % from pre to post-treatment and pre-treatment to follow-up respectively, whereas these figures were 46.2 and 53.8 % in the study by Lang et al. (
2012). The high rate of clinically significant change in the no treatment condition over the 1 week intervention period (46.2 %) is surprising, and it was not significantly different from that found in either the imagery or non-imagery condition. The emergence of the reduction in trait anxiety only at follow-up within the imagery condition in the current study could be due to the relative insensitivity to change of a trait scale, or it could be that changes in anxiety only emerged over a longer time period, once participants had had sufficient experience of deploying the newly trained bias in their daily lives, as has occurred in some other training studies (e.g. Browning et al.
2012).
The demonstration in the current study of the importance of the imagery instructions is an important translational step that builds on earlier experimental studies (Holmes et al.
2006; Holmes et al.
2008a,
2009b). In the current study, participants in the imagery condition were instructed to imagine themselves in the training scenarios, and participants in the non-imagery condition were not instructed to use any particular mode of processing. However, participants in both imagery and non-imagery conditions listened to the same positive training scenarios. It may at first seem surprising that positive information alone, that is, spending approximately 20 min every day for 1 week listening (with no imagery instructions) to hundreds of miniature stories with positive endings had no more impact on mood than engaging in no intervention whatsoever. However, this is consistent with our knowledge of the natural processing style observed in depression, and experimental studies comparing imagery to verbal processing.
In the current study, in the absence of instruction to use a particular mode of processing, participants in our non-imagery condition appeared to process the positive training materials in the verbal, ruminative style that characterises depression (Koster et al.
2011), and may contribute to difficulties in using positive memories to improve mood (Werner-Seidler and Moulds
2012; Joormann et al.
2007). Experimental studies in non-clinical samples have demonstrated that verbal processing of positive material does not improve mood or bias, and can even lead to deterioration of mood over a single session of positive CBM-I (Holmes et al.
2006, Holmes et al.
2008a,
2009b). We might therefore expect that, in the absence of instructions to use imagery, depressed individuals might use their natural (verbal) processing style and thus fail to gain any benefit from the repeated sessions of CBM-I. However, the importance of the imagery instructions on clinical outcomes over repeated sessions of positive CBM-I in a depressed sample had hitherto not yet been investigated. This study therefore extends the implications of the earlier experimental work to a clinical population, and complements the previous clinical study by Lang et al. (
2012). While the study by Lang et al. (
2012) demonstrated the importance of the consistently positive resolutions of the training material to be imagined, the current study demonstrated the importance of being required to
imagine the consistently positive training materials. Taking together the results from this study with those from the study by Lang et al. (
2012), we now have some initial evidence that it may be the
combination of the use of mental imagery with the consistently positive resolution of the training stimuli that accounts for the clinical impact of the imagery CBM-I paradigm on symptoms of depression, rather than either aspect of the training in isolation (cf. Hirsch et al.
2006; Holmes et al.
2009a).
While previous studies (Blackwell and Holmes
2010; Lang et al.
2012) have demonstrated effects of the imagery CBM-I on symptoms of depression and cognitive bias, this study is the first to our knowledge to provide evidence that engaging in repeated practice in generating mental images over the course of the CBM-I intervention could result in improvements in mental imagery ability, specifically increased vividness of visual mental imagery. This is potentially important, as depression is associated with reduced imagery vividness (Torkan et al.
2012), and in particular with reduced vividness for positive future events (Morina et al.
2011). At the other end of the spectrum, optimism, which may be seen as the polar opposite of the pessimistic thinking style associated with depression, is associated with increased vividness of positive future imagery (Blackwell et al.
2013). As being unable to imagine positive events in the future may contribute to depressed mood, increasing the vividness of this imagery may have useful clinical benefits in depression.
Although our measure of imagery vividness, the VVIQ, is a general measure of imagery vividness rather than of positive imagery, demonstrating that imagery vividness can be improved via repeated practice is an important step in understanding the potential mechanisms by which imagery CBM-I may have a therapeutic impact in depression. Interestingly, participants in the imagery condition did not show a significant increase in their score on the SUIS (Reisberg et al.
2003). Thus, the repeated practice in using imagery during the training sessions did not appear to generalize to a tendency to use (non-emotional) everyday imagery more outside of the training sessions. This is perhaps unsurprising and it was not an a priori hypothesis that this would increase; rather we were keen to determine which (of the many) aspects of imagery would be influenced during the training. It is worth noting that our results are consistent with the suggestion that the quality of imagery (vividness) may be a more important target for intervention in depression than frequency of use of (non-emotional) imagery per se. The tendency to use imagery may not in itself be adaptive or maladaptive, as it can amplify the affective impact of both negative (Holmes and Mathews
2005) and positive (Holmes et al.
2006) information, and in fact inducing a more image-based (concrete) mode of processing during a success experience does not result in greater improvement in affect compared to a more verbal (abstract) mode of processing in dysphoric individuals (Hetherington and Moulds
2013). Therefore specifically improving imagery vividness, rather than encouraging frequency of imagery use more generally, may be a useful aspect of this CBM-I. This will be interesting to explore further in larger studies and perhaps with better measures of imagery (cf. Pearson et al.
2013).
The results of the current study must be interpreted in the context of several limitations. It is important to bear the sample size (
n = 13 per group at post-treatment and
n = 8 per group at follow-up) in mind when interpreting the results from this study. As a first attempt (to our knowledge) to implement imagery CBM-I in this novel population, a small pilot study was appropriate in order to provide the initial evidence that could justify the time, resources, and participant burden of larger clinical trial in this new setting. However, this small sample size means that the results can only be interpreted as providing encouraging preliminary evidence. The results from pre-treatment to follow-up in particular must be interpreted with caution, due to the lack of outcome measurement in the no treatment group at follow-up, and the attrition in the imagery and non-imagery groups at this time point. The high rate of attrition at follow-up was mostly due to practical difficulties in attending another face-to-face assessment session, and thus future studies could enhance collection of follow-up data by having this data completed remotely, e.g. via online questionnaires or phone interviews. One concern when such promising clinical results are obtained with a small study relates to whether they can be replicated, as a small sample limits the generalizability of a study’s findings. We note that the results of this study are consistent with the two previous clinical studies (Blackwell and Holmes
2010; Lang et al.
2012), the experimental studies that preceded these (Holmes et al.
2006; Holmes et al.
2008a,
2009b), and a later randomized controlled trial (Williams et al.
2013). As the current study forms part of such a series of experiments, we can be more confident in the potential replicability of the results than if it was one promising study in isolation. Initial clinical translation studies like the current pilot study are not intended to provide conclusive proofs, but rather form an important step in a treatment development process and must be interpreted in this context. Some further cautions related to the sample size are noted below in discussion of several specific analyses.
The correlational analyses must be interpreted with caution due to the sample size, which limits statistical analysis of the mechanisms of change. However, the correlations suggest that for the sample as a whole, reduction in symptoms of depression was related both to reduction in negative interpretive bias, and to increase in imagery vividness over the 1 week from pre to post-treatment. Although it is not possible to draw conclusions about causal inference from these correlational data, they are at least consistent with the argument that the greater the extent to which whatever intervention (or no intervention) the participants engaged in modified their interpretive bias or imagery vividness, the more they experienced a reduction in symptoms of depression. The relationship between the putative active mechanisms of change in this CBM-I, imagery and interpretation, and clinical outcomes will be important to investigate more fully in larger samples and over longer time periods. It will also be important to investigate the potential interaction between the cognitive biases targeted (e.g. Everaert et al.
2012; Hirsch et al.
2006; Salemink et al.
2010; Tran et al.
2011). For example, studies in healthy volunteers (e.g. Holmes et al.
2006) have demonstrated that a single session of imagery CBM-I has an immediate impact on training a more positive bias and increasing positive mood. In order to further enhance the clinical potential of the CBM-I paradigm it will be useful to investigate the relative impact of these immediate effects of training sessions (e.g. change in bias, transient increase in positive mood) on longer-term clinical outcomes. The significant increase in one of the six outcome measures (scrambled sentences test) from pre-treatment to follow-up within the non-imagery condition would also need further investigation in larger samples.
Mental imagery represents a number of complex cognitive processes and plays numerous roles in daily functioning (Holmes and Mathews
2010). It will be useful for future studies to more fully characterise the impact on various aspects of mental imagery by using a more comprehensive range of measures (cf. Pearson et al.
2013), and imagery-based measures of cognitive bias (Berna et al.
2011). For example, as self-report measures of imagery such as used in this study may be subject to demand, it will be useful to investigate the effects of imagery interventions on performance-based measures of imagery ability. Further, it will be important to investigate whether the training has a differential impact on positive and negative imagery, and whether the fact that participants are required to generate field perspective imagery reduces the bias in depression to take an observer perspective (Nelis et al.
2013; Williams and Moulds
2007). Finally, it is important to note that another potential explanation for the superiority of the imagery condition in this study is that it required active generation of the positive outcomes, in the form of mental images, and it was this active generation rather than imagery per se that was the crucial difference between the two active conditions (cf. Hoppitt et al.
2010).
The development and dissemination of novel treatment approaches to help tackle the global health problem presented by depression requires that novel interventions translate from one country to another. This pilot study provides some preliminary evidence that the benefits of training positive imagination could possibly transcend national and cultural boundaries and encourages further investigation of the application of imagery CBM-I in novel populations. Further, it indicates the potential importance of mental imagery in translating the positive resolutions of the training scenarios into positive clinical outcomes in depression.